With SAS On Demand, the actual SAS process (which you access through the SAS Studio web application) runs on a Linux-based cloud server. You must use UNIX-style syntax for file and path names. UNIX does not have drive letters, but a single directory tree which starts at "root", symbolized by a slash. You "own" your home directory, and the system provides a shortcut to it, symbolized by a tilde. So if you created a directory/folder called mydata in your home folder in the Studio navigation tab, in your code you would use
libname mydata "~/mydata";
Also note that UNIX operating systems are case sensitive, mydata and Mydata would be different objects.
In the future, always post the complete log of code that fails by copy/pasting the text into a window opened with this button:
The log is our primary diagnostic tool, see Maxim 2.
... View more
Files in a ZIP archive need a two-step approach: first, define a FILENAME for the ZIP file then, in the DATA step you write to read the data, use the file reference and the name of the contained file as a member (see documentation of the INFILE statement)
... View more
Since SAS/ACCESS modules for relational databases make use of the native clients which have to be installed and configured on the SAS server, the SAS admins need to do this in cooperation with the DB admins. After they have done this, they can tell you which credentials to use.
... View more
The macro code (%SYSFUNC) is resolved before the DATA step is even compiled, much less executed. Therefore you cannot use DATA step variable values there.
Please describe in detail what you want to achieve, including the data that shall drive the code.
... View more
Why do you want to continue a job chain when there is an obvious problem that might cause your data to be irreversebly corrupted?
Anything else than a zero return code MUST not be allowed in automated execution.
... View more
Any non-zero code points to a problem and has to keep the scheduler from continuing with other jobs in the chain. To find out what caused the issue, read the log, and then correct your code or any other reason for the failure (like missing/incorrect input data).
... View more
Add the DISCRETEX option to the HEATMAP statement:
title "Injury Frequency by Year and Month";
proc sgplot data=sample;
heatmap
x=year y=MonthName /
colorresponse=COUNT
discretex
;
by InjuryType;
xaxis label="Year" values=(2020 2021 2022 2023 2024);
/* Character Y-axis preserves names; order guaranteed by sorting */
yaxis label="Month of Year" discreteorder=data;
run;
... View more
Wrap it in a macro:
%macro change_it(tabname);
data &tabname. (drop=value myperc_ label2);
set &tabname._;
mycount = input(scan(value , 1, ' ('), best.);
myperc_ = scan(value , 2, '()');
perc = compress(myperc_ , ' %');
label = label2;
run;
%mend;
%change_it(tab3) /* and so on */
... View more
You might have problems trying to make SAS prepare your coffee, but when it comes to date calculations, SAS has all tools needed, particularly the INTNX function:
data have;
input id $ date_01 :mmddyy10. date_02 :mmddyy10.;
format date: yymmdd10.;
datalines;
158 7/15/1994 9/9/2004
268 6/10/2003 8/21/2007
;
data want;
set have;
format want_01 want_02 yymmdd10.;
output;
want_01 = intnx("year",date_01,1,"b");
do while (want_01 lt intnx("year",date_02,0,"b"));
want_02 = intnx("year",want_01,0,"e");
output;
want_01 = intnx("year",want_01,1,"b");
end;
drop date_01 date_02;
run;
... View more
